Cost effectiveness of antimicrobial catheters in the intensive care unit: addressing uncertainty in the decision

Table 2 Monetary net benefits for a hypothetical evaluation comparing two novel treatments to standard practice

	Standard practice	Treatment A	Treatment B	Optimal choice
Simulation 1	140	150	160	B
Simulation 2	100	110	120	B
Simulation 3	110	100	100	Standard
Simulation 4	100	150	130	A
Simulation 5	130	120	110	Standard
Average expected net benefit	116	126	124	Standard/A/B = 40%/20%/40%

Results are expressed as monetary net benefits, each simulation is equally likely to be 'true'. Treatment A is associated with the highest expected net benefit (AUD $126), but because the distribution of monetary net benefits is skewed, it is preferred in only 20% of samples. Treatment A is therefore optimal, but the error probability associated with this choice is 80%. This probability is substantially higher than the 5% used for tests of statistical significance. The choice to remain with standard practice still carries a 40% probability of not returning the highest monetary net benefits and could be expected to incur economic costs of AUD $10 (AUD $126 minus AUD $116). The alternative with the highest monetary net benefit is the optimal decision, but that decision can be highly uncertain.

ISSN: 1364-8535